The gold did not run out — the owners did: Duarte da Silva's devastating case against SA's mining abandonment

South African gold output has fallen by roughly nine-tenths from its peak while the dollar price has multiplied more than twenty-fold — from a 1999 low near $255 to a January-2026 record above $5,500, before easing back toward $4,000 by mid-year. The decline of the deep mines was real and, on its own terms, correctly forecast. But the figure that wrote off the resource — a reserve calculated when gold traded near nine hundred dollars — was an economic boundary, not a geological one, and at today’s price it no longer binds. The houses dismantled themselves; the shafts were left to flood; the dumps were left full; and the men who came back for the gold were met with soldiers rather than a licence. The indictment is not that the gold ran out. It is that it did not, and the country abandoned it anyway..By Dr Duarte de Silva.1.The chargeStrip the courtroom imagery from the plate above and the accusation underneath it is precise. South Africa did not lose its gold industry to an act of God or a vein that pinched out. It lost it to a sequence of decisions — by the mining houses that built the deep levels and then declared them somebody else’s problem, and by a state that let six thousand shafts fall ownerless rather than govern what was left in them. The banner on the wall reads not a mistake, a choice, and the distinction is the case: an accident can be forgiven, where a deliberate choice has to be answered for.The choice has a shape that recurs in every part of this paper. Profit was taken when the rand-gold margin was wide and the easy ore was near surface. Liability was shed — to a smaller operator, to a politically connected shell, to a liquidator, finally to the taxpayer — the moment depth, grade and the pumping bill turned the same ground from an asset into a cost. Privatise the upside, socialise the flooded void. That pattern is not a footnote to the South African gold story; it runs through the middle of it.An accident is forgiven. A choice is judged. The banner reads: not a mistake — a choice.2.The men on the aeroplanesWalk the floor of any mining investment conference in 2026 and follow the money with your eyes. The exploration budgets are on aeroplanes — to West Africa, where a junta in Bamako jailed a chief executive and seized three tonnes of gold by helicopter; to the eastern Congo, where a war is on; to Nevada, to Western Australia, to anywhere a geologist can be flown to look for an orebody that has not yet been found. Capital is doing what capital does: searching the world for gold.Now stand still and look down. Beneath that conference floor, beneath the city it sits in, lies the Witwatersrand Basin — the source of roughly forty per cent of all the gold ever mined, and, on peer-reviewed estimate, still holding some 48,100 tonnes in the ground, with a further 1,600 tonnes on surface in the tailings. Twenty minutes from where I write, in Boksburg, the East Rand Proprietary Mines closed in 2008. Over one hundred and twelve years it produced forty-three million ounces; when it stopped it left sixty-three million ounces behind — more than it had ever taken out. It did not stop because the gold ran out. It stopped because a pump bill of about a hundred and fifteen million rand was more than a company losing money at $870 gold could carry.So the question is simple, and it should silence a room: why is anyone on a plane? The gold that capital is crossing the planet to find is documented, quantified and abandoned beneath the airport it flew out of. The answer is not geological. It is that the gold here has been filed under a single word — exhausted— and that word has proved more powerful than the orebody it describes. The rest of this paper takes that word apart, and then asks who benefits from leaving it intact, and what leaving it intact has cost.Why is anyone on a plane? The gold that capital crosses the planet to find is abandoned beneath the airport it flew out of. E X H I B I T A · T H E D I V E R G E N C E3.Output collapsed through the greatest gold bull market in historyBegin with the two lines, because their crossing is the fact everything else has to explain. The country produced about a thousand tonnes of gold in 1970, two-thirds of the non-Soviet world’s output. By 1980 it was near 675 tonnes; by 2000 around 431; in 2024 and 2025 roughly 90 — the lowest level since the nineteenth century. Over the same span the gold price did the opposite, climbing from a 1999 low near $255 to a record above $5,500 on 28 January 2026 — before a nine-per-cent single-day drop and a long retracement left it near $4,000 by mid-year. The chart below traces the run-up and the pullback alike; the trajectory, not the last tick, is the point..Plotted against each other, the divergence is unmistakable, and it is the whole of the problem. For most of economic history a shrinking industry tracks a falling price. The South African gold sector did the reverse: it contracted while the thing it sold became more valuable every year. That is not a normal depletion curve. It is what it looks like when an industry walks away from ore it can increasingly afford to lift, and the widening space between the lines is the revenue per tonne the country simply stopped capturing..The widening gap between the two lines is the revenue per tonne the country stopped capturing — not because the gold was gone, but because the owners had left. E X H I B I T B · T H E R O L L - C A L L O F T H E A B A N D O N E D1.The houses dismantled themselves — mine by mineThe fall in the chart is not one event but the sum of dozens of closures, and the closures were not random. The corporate map of South African gold was redrawn almost completely in twenty-five years, and the defining move throughout was capital walking away from the deep-level mines it had built. What follows is the record — not two cases, but the pattern across the whole field.The majors: who left, who stayedAnglo American. Anglo consolidated its gold interests into AngloGold in 1998, merged it with Ghana’s Ashanti Goldfields in 2004 to form AngloGold Ashanti, then sold down steadily and exited gold entirely by 2009. The Free State assets went to African Rainbow Minerals and Harmony. In October 2020 it sold its last South African mine — Mponeng, the deepest in the world — together with Mine Waste Solutions to Harmony for about R4.4 billion, ending more than a century of Oppenheimer-era gold mining in the country.Gold Fields. In 2013 Gold Fields unbundled its mature South African shafts — Kloof, Driefontein and Beatrix — into a new vehicle, Sibanye Gold, keeping only the mechanised South Deep for itself. Sibanye (now Sibanye-Stillwater) went on to absorb Cooke, Wits Gold and Burnstone, then diversified into platinum. Gold Fields, like Anglo before it, turned its growth offshore — but unlike Anglo it kept a foot in the basin, holding South Deep, the mechanised mine it still works and expects to outlast every other gold mine in the country.Harmony Gold. Harmony became the receptacle for what the majors discarded and is today the country’s largest gold producer, with group output of about 1.47 million ounces — roughly forty-six tonnes — in the year to June 2025, down five per cent on the year before. The contrast is the whole indictment in miniature: the original owners treated South African deep gold as a liability to be offloaded; a smaller, lower-cost operator treated the very same assets as a business. The ore did not change. The willingness to mine it did.The mines that went dark — and whyBeneath the corporate reshuffling, individual mines closed for concrete and repeating reasons. The same cluster of killers appears again and again — depth, declining head grade, the electricity bill, labour cost and unrest, safety, and above all water: acid mine drainage and the cost of keeping deep workings dry. What is almost never on the list is running out of gold in the ground.East Rand Proprietary Mines (ERPM), Boksburg — 1893 to 2008. One of the founding mines of the Witwatersrand and the deepest on earth until 2008, at roughly 3,585 metres. As the last working mine in the Central Rand basin it inherited the dewatering burden of every shaft that had closed up-gradient of it; the water-management bill, not exhaustion, is what finally ended underground operations after 112 years. It had produced some forty-three million ounces and left an estimated sixty-three million behind — more gold than it ever lifted. The pump bill that broke it, about a hundred and fifteen million rand, was a fraction of the value it sat on, but more than a company losing money at $870 could carry. Surface retreatment of its dumps continued; the underground rights were sold off in 2019. The void it left is now part of the acid-water problem the state pays to pump.TauTona and Savuka (Western Deep Levels), Carletonville — 1962 to 2018. TauTona, the “great lion,” was extended to 3.9 kilometres in 2008 and briefly held the title of deepest mine in the world. Rising seismicity, falling grade and the sheer cost of working at that depth ran its life down; AngloGold Ashanti placed it and the neighbouring Savuka section into orderly closure in 2018, folding the remaining reserves into Mponeng. Savuka had already gone onto care and maintenance in 2017, kept alive only as a pumping shaft to keep the others dry. The deepest engineering achievement on the continent, retired not for want of ore but for want of margin.Blyvooruitzicht, Carletonville — 1937 to 2013. Once among the richest mines on earth. A slumping rand-gold margin, declining grade, rising cost and labour disputes forced it into liquidation roughly fourteen years ahead of its planned life. The owners, DRDGold and Village Main Reef, walked away; the village collapsed into unemployment and acid mine drainage. It is the textbook case of disorderly closure — the moment a community discovers that a mine’s liabilities outlive its dividends, and that no one has been left holding them. It returns later in this paper for a second reason: in 2022 it came back to life.Grootvlei and Orkney (Pamodzi, then Aurora) — 2009 to 2011. Driven under in the price-stress years and handed to the politically connected Aurora consortium, whose directors included Zondwa Mandela and Khulubuse Zuma. Aurora stripped the assets rather than ran them; the pumpsfailed, the water rose, and two operating mines were destroyed. The courts later found the directors personally liable for the wreckage. It is the purest example in the record of the abandonment dressed up as a rescue.Buffelsfontein / Stilfontein, North West — ceased 2013. A major producer in its day, shut and left. Its unsealed shafts became the stage for the events of 2024 that give this series its sharpest edge — a subject this paper returns to in Exhibit G. The mine’s closure is unremarkable in the roll-call; what happened in the hole afterwards is not.Behind the named mines stand the goldfields themselves — the Central and East Rand exhausted at workable depth and beaten by pumping costs; the Far West Rand, Klerksdorp, the Free State around Welkom and Evander maturing out the great 1952–65 boom. The list of survivors is short, and shrinking. The point is not the obituary of any single shaft. It is that scores of them closed while the price of what they produced was climbing toward a record — and that the response, every time, was to leave.The ore did not change. The willingness to mine it did. The majors called the deep levels a liability; a smaller operator called the same rock a business. E X H I B I T C · T H E G O L D T H E Y L E F T B E H I N D2.The 36,000-tonne myth — and the reserve drawn at a $900 priceFor years South Africa claimed first place in world gold reserves, on a suspiciously round and, after 2001, unchanging figure of 36,000 tonnes — about 40 per cent of the global total. The number was a reserve base lifted from government yearbooks, not a demonstrated, economically mineable reserve, and the distinction decides everything. Three quantities get blurred together: the geological endowment, all the gold the Witwatersrand ever held; the resource, the part identified and quantified; and the reserve, the fraction that can be economically extracted at the time of determination. Only the last depends on price — and price is exactly what has moved.A reserve is a line drawn through an orebody at the price of the day. Draw it at $300 gold and most of the rock is waste; draw it at $4,000 and most of that same rock is ore. On a standard cut-off calculation, holding cost flat, the grade at which a deep-level tonne pays falls from around eleven grams at $300 to under a gram at today’s price. But that figure flatters itself. It assumes nothing has happened to the cost of mining since gold was $300 — a world of roughly 1999, before a quarter-century of power tariffs, wages and depth all climbed. They climbed steeply. Put the cost back in — call it two to three times the dollar cost per tonne of that era — and the honest mineable line settles nearer two grams a tonne, not one. Still a fraction of the five to eleven grams that once defined ore, and the point is unchanged: a great deal of what was waste at $300 is ore at $4,000. The price has made that journey. The mine plans have not moved at all — the operators are still mining to lines drawn for a world that ended around 2005.The most rigorous public deflation of the 36,000-tonne claim is Hartnady’s 2009 analysis in the South African Journal of Science. Rather than trust a yearbook, he applied Hubbert-style depletion modelling — the curve-fitting method used for oil — to the full production history, projecting an ultimate recoverable resource of about 53,800 tonnes. Subtract what had already been mined and the residual reserve came to roughly 2,950 tonnes: less than a tenth of the official figure. His forecast that output would fall permanently below 100 tonnes a year within the decade was simply correct, and the USGS has since cut its South African reserve estimate toward 5,000 tonnes. On its own terms, the bear case was vindicated.But look at what that number actually measures. A Hubbert reserve is not a geological constant; it is an economic limit that depends entirely on price. When Hartnady wrote, gold averaged under $900, and his figure implicitly answered one question: how much gold is worth lifting from these depths at roughly $900? Move the price to $4,000 and the boundary moves with it — waste rock becomes ore, and the same mathematics that wrote the epitaph quietly retracts it. The production forecast held. The reserve framework behind it broke.The evidence is in the miners’ own filings. Mponeng — the deepest mine on earth — carries a mineral resource of 23.5 million ounces against a booked reserve of roughly 2.1 million. The twenty-one-million-ounce gap is gold the company owns and does not count, because no one has redrawn the reserve line down the reef at a price that now plainly exists. Harmony’s own chief executive has called the reserves below the current four-kilometre level “massive” — while the same company returns roughly half its free cash flow to shareholders as dividends and buybacks. A farmer who sells his seed corn in a good year is not being disciplined; he is eating next year’s crop to flatter this year’s accounts.What the depletion curve never counted is the surface. The Hubbert model measures one thing, primary underground production, and is structurally blind to the dumps. Three separate, additive categories sit outside it. The underground endowment of the Witwatersrand Basin is on the order of 48,100 tonnes. The Witwatersrand surface tailings hold about 1,600 tonnes. And separately — additive to that tailings figure, not a fraction of it — Chingwaru and colleagues (2023, Scientific Reports) quantified roughly 420 tonnes of “invisible” gold locked in pyrite and arsenian pyrite, discarded by conventional cyanide retreatment that recovers only the free-gold fraction. That single overlooked sulphide-hosted resource rivals one of the largest operations on the planet — and it is already on surface, already mined, requiring no shaft, no depth, no dewatering..This is why the one expanding part of the South African gold industry is not underground at all. Surface retreatment — DRDGold’s Ergo, Pan African Resources’ Elikhulu and the Mintails West Rand assetsis the segment whose economics improve as the orebodies underground get poorer, because a higher gold price lowers the grade you can profitably treat. It is the structural inverse of the deep mine. As price rises, the dump becomes richer to work and the shaft becomes harder; the industry the houses abandoned is being quietly rebuilt on the waste they left.The pessimists were right about the mines and wrong about the gold. The reserve that wrote off South African gold was an artefact of a $900 price; the question is no longer whether the gold is there, but who gets it out.E X H I B I T D · T H E M O D E L T H A T P R I C E S D E C L I N E1.Why the investor cannot back itThe asset manager has a different failure, and it is anchoring — operating here with a particular, compounding cruelty. The models that value these companies still run $2,000 to $2,500 long-term gold against a spot price near $4,000. The political-risk discount applied to South Africa is frozen at a calibration set when gold was $1,500. Both are defended as prudence. Neither has been updated, because being wrong with the consensus is survivable and being right alone is not.There is a second failure here, specific to this asset, and more damning than anchoring. The upside is in the resource, and the discounted-cash-flow model only prices the reserve. Reclassification — the single largest source of value in the entire South African gold complex — is structurally invisible to the instrument used to value it. The analyst takes the reserve as given, prices a wasting asset running down to depletion, and is mechanically incapable of seeing the far larger asset that redrawing the reserve line would create. The model cannot price the very thing the higher price has created. It values the run-off, and reports the result as though it were objective.And where the proof is already audited and public, it is ignored anyway. DRDGOLD and Pan African Resources earn surface-retreatment margins at today’s gold price that should, on any honest reading, have re-rated the entire secondary-recovery category. They have not, because the models pricing them are still anchored to a gold price that no longer exists. This is the slow car on the motorway, and the slow car does not believe it is a hazard. It believes it is being careful, while the projects that are not funded and the assets that are not developed pile up in the lane behind it.The model cannot price the very thing the price has made possible. So it prices decline, and calls the number objective. E X H I B I T E · T H E L A S T S H A F T2.When did this country last sink a deep shaft?Here is a question to put to anyone who tells you South African gold is a sunset industry: when did this country last sink a deep gold shaft? The answer is exact, and it is damning. The last major deep-level gold shaft sunk in South Africa was the Twin Shaft complex at South Deep. Sinking began in 1995; it was commissioned in 2005. Moab Khotsong, the only other contender, came into production in 2003. Since then — nothing. A country that sank shafts continuously for one hundred and ten years, that built the deepest mines in the history of the species — eight of the ten deepest on earth are South African gold mines — simply stopped. Two decades have passed in which not one new deep gold shaft has gone into the ground.In 2026 West Wits opened Qala Shallows, hailed as the first new South African gold mine in fifteen years. Read the detail: it is a decline to eight hundred and fifty metres on Central Rand remnant reefa shallow pick-up of ground left behind, not a shaft. It is the exception that measures the silence around it. This is an industry that has forgotten how to begin. It remembers only how to close — and, increasingly, how to be closed.Eight of the ten deepest mines on earth are South African. The last deep shaft this country sank was commissioned in 2005. It has forgotten how to begin. E X H I B I T F · T H E S T A T E T H A T C L O S E S H O L E S3.Closing holes instead of opening themIf the miners are paralysed and the investors anchored, the state is something worse: absent where it should build, and brutally present where it should not be. Begin with the numbers it controls. Real exploration expenditure in South Africa was R6.2 billion in 2006. It was R738 million in 2025. The spending that finds the next mine has collapsed by roughly seven-eighths in nominal rand, across the very two decades in which the rand gold price multiplied many times over. South Africa now attracts under one per cent of global exploration spend, against the five per cent the responsible minister says he is targeting — and the five per cent the country actually commanded a generation ago.The Fraser Institute’s 2025 survey ranks South Africa 64th out of 68 mining jurisdictions on policy alone — beneath Guinea, beneath Burkina Faso, beneath Mali. Beneath, that is, the very destinations to which South African capital is flying to look for gold. And the most basic instrument of a functioning minerals economy — a cadastre, the online map of who holds the right to what — still does not work. A new system was promised live in June 2025. It is operational in exactly one province, the Western Cape, chosen because it has the least mining of any. The national rollout has slipped to March 2027. The state cannot, at present, draw the map of its own mineral wealth.Then there is the matter of what the state does do in the goldfields, which is the part that should be read twice. The single largest state intervention in South African gold this decade is a police operation named Operation Vala Umgodi. The name is Zulu. It means: close the hole. The flagship action of the government of the richest goldfield ever found is an operation to seal mines shut — not to open them, not to secure and re-permit them, not to put people to work in them. To close them. With people still inside.Exploration spend, R6.2 billion in 2006, was R738 million in 2025. The state’s flagship action in the goldfields is named, in plain Zulu, close the hole. E X H I B I T G · T H E C L O S U R E T H A T N E V E R C O M E S4.The certificate that is never signedThere is a single document at the centre of all of this, and almost no one outside the industry has heard of it. Under Section 43 of the Mineral and Petroleum Resources Development Act, a mine is not closed when it stops producing, when the workers are paid off, or when the headgear is sold for scrap. It is closed — legally, finally — only when the Minister issues a closure certificate. Until that certificate exists, the holder of the mining right remains responsible in law for the ground, the water and the rehabilitation. The certificate is the moment the liability is meant to end. It is also the moment that, in South Africa, almost never comes.This is not a technicality; it is the whole mechanism. The country’s derelict and ownerless mines are not defined by being old, or empty, or forgotten. They are defined — in the Department’s own wordsas mines for which no closure certificate was ever issued and whose owners can no longer be traced. The abandonment and the missing certificate describe one situation from two sides. By the Department’s 2025 count there are roughly 6,100 of them, alongside some 1,170 unsealed, unsafe openings, and fewer than one in a hundred has been rehabilitated. What is missing is not the law. The law is clear, and it is strict. What is missing is the signature at the end of it — and the enforcement that signature was supposed to represent.Ask why the certificate is so rarely granted and you reach the deadlock that runs underneath the whole goldfield. A closure certificate is the state certifying that a mine has been made safe and that residual liability can be released. But the residual liability here is acid mine drainage — water that will decant, and have to be pumped and treated, in effect forever. No official wants to sign a document that handsthe state a perpetual, unquantifiable cost; and the environmental law has moved the other way, attaching liability to the company in perpetuity even after a certificate is issued. So the certificate that was meant to end the obligation now ends nothing, and is, rationally, almost never sought or signed. The mine is left in permanent limbo: neither operating nor closed, its owner neither present nor formally released, its liabilities sitting in the dark with no name attached.Into that limbo the owners leave by other doors. A company that cannot obtain a closure certificate does not need one in order to walk away. It can place the mine on indefinite care and maintenance and let it rot. It can sell it down a chain of ever-smaller operators until the last buyer is a shell with no balance sheet — the Aurora model. Or it can enter business rescue or liquidation, where environmental obligations rank behind the banks and routinely evaporate. Blyvooruitzicht’s owners left through liquidation; Grootvlei’s and Orkney’s through a consortium that stripped them. None obtained a closure certificate. None needed to. The front door marked responsibility is permanently locked; the abandonment goes out the back.Behind the unsigned certificate sits an underfunded one. Every mine is required to set aside financial provision for its eventual rehabilitation, and the gold sector’s provisions have for years been widely found to fall far short of true closure cost — understated while the mine is alive, because a fully funded closure liability is a fully visible one. When the company then fails, or simply leaves, the gap between what was provisioned and what closure actually costs does not vanish. It becomes the roughly R49 billion in rehabilitation the state cannot pay. The under-provisioned fund and the un-issued certificate are two halves of one manoeuvre: book the smallest liability the rules allow while mining, and make sure the document that would crystallise the rest is never signed.This is the machinery beneath the human story that follows. Stilfontein was not an aberration; it was the system working precisely as it was built to work. A mine ceased operating in 2013. No closure certificate was issued. The shafts were never sealed and rehabilitated to the standard the law requires. The owner was gone, the certificate unsigned, the ground left open — a legally defined nothing, into which came the men the state would later try to flush out with hunger. The certificate that was never signed and the bodies later carried up the shaft are the same event, recorded first in an administrative file and then in a mortuary.A derelict mine is not one that ran out of gold. It is one for which the closure certificate was never signed — and the owner can no longer be found. E X H I B I T H · T H E D A M A G E A N D T H E D E A D5.The bill the owners did not payWalking away is never free; it only moves the cost — and the first form the cost takes is water. The deep voids the houses stopped pumping fill with water that turns acid on the pyrite, and that water rises — toward aquifers, rivers and, in the Central Rand, the surface of Johannesburg itself. The pumping that the mines once funded as an operating cost is now a public liability, run at places like the Central Basin by a state agency because the owners are gone. Acid mine drainage is the abandonment made visible: a poison the companies created, decanted onto a public that never saw the dividends.And the water tells the story in slower motion. The Witwatersrand was never a set of separate mines. Below ground it is one connected body of water, divided into sub-basins by nothing more than pillars of rock and a handful of concrete plugs, and held down only by pumping. For as long as every mine pumped, the system stayed dry. The failure was built to spread. As each mine closed and switched off its pumps, the water it had been lifting did not vanish; it moved next door, to whoever was still pumping. By 2004 the Central Rand was flooded above its 24 Level while the neighbouring workingsstood full to 50, and within a year the water was spilling from one sub-basin into the next, forcing the installation of high-pressure bulkhead plugs to hold it back.ERPM was the last mine still pumping in the Central Rand, and so it inherited the water of every mine that had quit before it. In 2008 two of its workers died from gas driven up the shafts by the rising flood, and the Department of Water Affairs ordered the workings stopped. The pump bill that finally broke the mine was never the cost of one mine’s water — it was the cost of a whole basin’s, every upstream owner having already walked away and left the last one standing to drown. Grootvlei and Orkney went the same way by other means: handed to a consortium that stripped them, their pumps failed and the water rose. Switch the pumps off in a boardroom or leave them to rot — the basin does the same work either way, and the state is left the bill.The administrative limbo of the previous exhibit has a body count. Into the unsealed, uncertified workings of Stilfontein — abandoned in 2013, never properly closed, in the words of the Bench-Marks Foundation — came the zama zamas: desperate men, many undocumented, many co-opted by syndicates run by former mine employees, sent underground for months to scrape a fraction of the gold that made this country rich. This is where the abandonment stops being a balance-sheet story.From August 2024 the state’s answer arrived. At the abandoned Buffelsfontein shafts the food and water were cut off and the openings sealed, to flush the miners out. At least seventy-eight came up dead; some counts run to a hundred. One thousand two hundred and four surfaced alive. The Human Rights Commission later heard that twenty-seven of those underground were children, and that most who surfaced were trafficking victims. At Sabie in December 2024 another such operation brought a hundred and fifty-three to the surface and three to the morgue; at Krugersdorp in 2022, eight women were gang-raped at an abandoned mine — the human cost of unsecured shafts written in the plainest possible terms. The government puts the cost of illegal mining at sixty billion rand in 2024 alone. The operation at Stilfontein was described, officially, as a success.Now set two facts side by side. The gold those men died for is real — the same gold that is mispriced, abundant and recoverable. And the state’s response to a desperate population standing on a world-class orebody was to fence the orebody and seal the population inside it. Treat the symptom, the man in the hole. Leave the cause — the abandoned, unclosed, uncapitalised, world-class orebody — exactly where it was.The zama zama is not the crime. The zama zama is the verdict. E X H I B I T I · T H E W A Y B A C K6.What it would actually takeThe case against is always the same, and it is wrong in a specific way. People hear “restart the goldfields” and picture a decade and billions of rand sunk into a new deep shaft — the very thing this country last did in 1995 and has been afraid to attempt since. That is the most expensive, slowest and last tier of the work. The mistake is to begin there. Begin at the other end, where the cash is fastest and the risk is lowest, and let the early tiers fund the later ones.Tier one is surface — the tailings. Sixteen hundred tonnes of gold sit in the surface dumps, and, separately, about four hundred tonnes more of sulphide-locked “invisible” gold the old plants could never capture, now recoverable with technology that exists today. The two are additive, not the same tonnes counted twice. No shaft, no depth, no dewatering, and economics that improve as the price rises, because a higher price lowers the grade worth treating. DRDGold and Pan African already prove the model at scale. This is months to cash, not years — and it is where a restart begins.Tier two is shallow remnant and decline reopening. The Qala model. The Blyvoor model — a mine declared dead in 2013, producing again from 2022, with fifteen hundred people back at work. Existing headgear, existing infrastructure, the reserve line simply redrawn at the $4,000 cut-off to take in the secondary reefs and pillars that were waste at $300. Twelve to thirty-six months, not a decade.Tier three is the flooded basins. ERPM and the Central and East Rand. Here the binding constraint is the same water that cascaded the mines shut in the first place — and that water is already a state liability, the acid-drainage pumping bill paid right now, in perpetuity, for no return whatsoever. Pair remediation with recovery and the pumping you are already funding becomes the access you need. Sixty-three million ounces at ERPM, set against a pumping cost once quoted in the low hundreds of millions of rand, at $4,000 gold. The cost centre is the orebody.Only after those three do you reach the deep shaft — and by then the cash flow from the first three is what funds it. As for the capital, it already exists: the miners are generating record free cash flow and returning half of it to shareholders. Redirect a fraction of that from buybacks to reclassification and reopening, and the sector funds its own restart. The institutions need only price the resource rather than the reserve alone. And the state must do the cheapest thing of all — not subsidy, but certainty: a cadastre that works, secure and unambiguous title, an acid-drainage partnership instead of an acid-drainage liability, and an end to discretionary licensing. De-risk; do not fund.The architecture that turns this ladder into policy — the compact that trades regulatory certainty for binding commitment, and converts a stranded liability into investment, employment and remediation— is the subject of Paper 5, The Licence to Dig. What matters here is only that the ladder exists, that it is climbed cheapest rung first, and that the capital and the gold to climb it are already on the books.You do not need the whole endowment. A single-digit percentage of 48,100 tonnes, reclassified at today’s price, is a multi-decade industry and tens of thousands of formal jobs — the same men, in hard hats, paying tax, instead of dying in sealed shafts. E X H I B I T J · T H E C H A R G E S H E E T7.Three failures, and a question of lawThree failures, one asset. The miner who would rather shrink the company than redraw a line on the orebody he already owns. The investor who prices the wasting asset with precision and is structurally blind to the far larger one that reclassification would build. The state whose exploration budget has collapsed by seven-eighths, whose map of its own minerals works in a single province, and whose flagship action in the goldfields is named, in plain translation, close the hole. And beneath all three of them, mapped and abandoned, flooding and looted, the richest goldfield ever found — killing the very people the state sends police to seal inside it.For twenty years this has been called failure. The word is too soft, and it may also be wrong in law. Take the abandonment first. To walk away from a mapped, world-class orebody and leave it to flood, to leach acid into the water table, and to kill, is the kind of act the law is slowly learning to name.Ecocide is not yet a crime under the Rome Statute, but the direction of travel runs one way: France has written it into its own penal code, and in 2026 the International Criminal Court issued a policy treating grave environmental destruction as a means by which existing crimes can be committed — with senior decision-makers, as individuals rather than the companies they ran, squarely in scope.Stilfontein is sharper still. Sealing the shafts and cutting off the food and the water to drive a trapped population to the surface is not a figure of speech for a crime. It has the shape of one. The deliberate,systematic deprivation of a civilian population is the conduct the law of crimes against humanity was written to reach, and a charge of that order does not require a new statute — only a prosecutor willing to read the facts as what they are.And the comfortable assumption — that the houses which dismantled this industry and left the country are now beyond reach — is false. The dismantling is a matter of record: Anglo American folded its gold into a separate company and was out of the metal by 2009, selling the deepest mine on earth to a smaller rival in 2020; Gold Fields unbundled its ageing South African shafts in 2013 and kept only its one mechanised mine. That the original owners treated the country’s deep gold as a liability to be offloaded does not place them beyond reach. In Lungowe v Vedanta the United Kingdom Supreme Court allowed more than eighteen hundred Zambian villagers to sue a London-domiciled parent company, in London, for pollution caused by its subsidiary abroad; Okpabi v Shell confirmed the principle. Many of the groups that hollowed out South African gold still answer to a parent listed in London or Johannesburg, reachable in its own courts. And the industry has already been brought to book once, in living memory: for the dust that gave tens of thousands of miners silicosis, six of the largest gold companies now fund the Tshiamiso Trust — at least five billion rand, their liability uncapped, paying out to men who worked the reefs as far back as 1965. Dust inhaled underground was actionable. Acid water, sealed shafts and a body count are not a lesser case.This is not an indictment drawn by a prosecutor. It is the charge sheet a prosecutor could start from. The facts are mapped, audited and public. What is missing is someone with the standing to file them.Dust inhaled underground was actionable, and six houses now fund the Tshiamiso Trust. Acid water, sealed shafts and a body count are not a lesser case. T H E V E R D I C T8.Not a mistake. A choice.Put the exhibits together and the case closes itself. The deep mines really are in terminal decline — depth, grade, electricity, labour and water make new conventional underground gold mines uneconomic almost regardless of price, and on that narrow point the pessimists were right. But that was never the accusation. The accusation is that an industry took the profit, shed the liability, left the shafts to flood and the dumps full, and then watched a state seal holes and bury men rather than govern the wealth still in the ground. The reserve that justified the epitaph was drawn at $900. At $4,000 it is void, and the binding constraint has shifted from geology to recovery technology and processing economics.The crime was the abandonment — the decision to walk away from the richest goldfield ever found and leave it to flood and to rot. The verdict is written in the acid water, the derelict shafts and the count of the dead at Stilfontein. The sentence does not have to be more of the same. The gold is still there, on surface and at shallow depth, in quantities the 2009 models never counted; the people are still there; and the plants that can recover it are already built, listed and audited. What is missing is not the resource. It is the decision to stop abandoning it.The geology did its work three billion years ago and has not failed once since. Everything that has gone wrong in this story is human, recent, and reversible. The only open question is the one the conference floor keeps refusing to ask out loud: whether anyone with the standing to act will look down, at the basin under their feet, before the next aeroplane to Bamako boards..Sign up for your early morning brew of the BizNews Insider to keep you up to speed with the content that matters. The newsletter will land in your inbox every morning on weekdays. Register here.Support South Africa's bastion of independent journalism, offering balanced insights on investments, business, and the political economy, by joining BizNews Premium. Register here.If you prefer WhatsApp for updates, sign up to the BizNews channel here.

The gold did not run out — the owners did: Duarte da Silva's devastating case against SA's mining abandonment

Related Stories

Duarte da Silva: South Africa's gold is still there. The architecture to unlock it isn't — yet

SA firms caught in crossfire as anti-immigration protests spark continental backlash

Last call: Why SA's discarded gold assets are now the world's most compelling trade

Shirley de Villiers: State of the city - thoughts and prayers for Joburg